DOCSLIB.ORG

CONCENTRATION INEQUALITIES for MEAN FIELD PARTICLE MODELS 3 As the Gibbs Measure, Defined By

Total Page：16

File Type：pdf, Size：1020Kb

Download full-text PDF Read full-text
CONCENTRATION INEQUALITIES for MEAN FIELD PARTICLE MODELS 3 As the Gibbs Measure, Defined By The Annals of Applied Probability 2011, Vol. 21, No. 3, 1017–1052 DOI: 10.1214/10-AAP716 c Institute of Mathematical Statistics, 2011 CONCENTRATION INEQUALITIES FOR MEAN FIELD PARTICLE MODELS By Pierre Del Moral and Emmanuel Rio INRIA Bordeaux-Sud-Ouest, and INRIA Bordeaux-Sud-Ouest and Universit´ede Versailles This article is concerned with the fluctuations and the concen- tration properties of a general class of discrete generation and mean field particle interpretations of nonlinear measure valued processes. We combine an original stochastic perturbation analysis with a con- centration analysis for triangular arrays of conditionally independent random sequences, which may be of independent interest. Under some additional stability properties of the limiting measure valued pro- cesses, uniform concentration properties, with respect to the time parameter, are also derived. The concentration inequalities presented here generalize the classical Hoeffding, Bernstein and Bennett in- equalities for independent random sequences to interacting particle systems, yielding very new results for this class of models. We illustrate these results in the context of McKean–Vlasov-type diffusion models, McKean collision-type models of gases and of a class of Feynman–Kac distribution flows arising in stochastic engineering sciences and in molecular chemistry. 1. Introduction. 1.1. Mean field particle models. Let (En)n 0 be a sequence of measur- ≥ able spaces equipped with some σ-fields ( n)n 0, and we let (En) be the set of all probability measures over the setE E ≥, with n 0. WeP consider a n ≥ collection of transformations Φn+1 : (En) (En+1), n 0, and we denote by (η ) a sequence of probabilityP measures→ P on E satisfying≥ a nonlinear arXiv:1211.1837v1 [math.PR] 8 Nov 2012 n n 0 n equation≥ of the following form: (1.1) ηn+1 =Φn+1(ηn). Received April 2010. AMS 2000 subject classifications. Primary 60E15, 60K35; secondary 60F99, 60F10, 82C22. Key words and phrases. Concentration inequalities, mean field particle models, mea- sure valued processes, Feynman–Kac semigroups, McKean–Vlasov models. This is an electronic reprint of the original article published by the Institute of Mathematical Statistics in The Annals of Applied Probability, 2011, Vol. 21, No. 3, 1017–1052. This reprint differs from the original in pagination and typographic detail. 1 2 P. DEL MORAL AND E. RIO These discrete time versions of conservative and nonlinear integro-differential type equations in distribution spaces arise in a variety of scientific disciplines including in physics, biology, information theory and engineering sciences. To motivate the article, before describing their mean field particle inter- pretations, we illustrate these rather abstract evolution models working out explicitly some of these equations in a series of concrete examples. The first one is related to nonlinear filtering problems arising in signal processing. Suppose we are given a pair signal-observation Markov chain (Xn,Yn)n 0 on ≥ some product space (Rd1 Rd2 ), with some initial distribution and Markov transition of the following× form: P((X ,Y ) d(x,y)) = η (dx)g (x,y)λ (dy), 0 0 ∈ 0 0 0 P((X ,Y ) d(x,y) (X ,Y )) = M (X , dx)g (x,y)λ (dy). n+1 n+1 ∈ | n n n+1 n n+1 n+1 In the above display, λn stands for some reference probability measures d2 on R , gn is a sequence of positive functions, Mn+1 are Markov transitions d1 from R into itself and finally η0 stands for some initial probability mea- sure on Rd1 . For a given sequence of observations Y = y delivered by some sensor, the filtering problem consists of computing sequentially the flow of conditional distributions defined by η = Law(X Y = y ,...,Y = y ) n n| 0 0 n n and b η = Law(X Y = y ,...,Y = y ). n+1 n+1| 0 0 n n These distributions satisfy a nonlinear evolution equation of the form (1.1) with the transformations Φn+1(ηn)(dx′)= ηn(dx)Mn+1(x,dx′) and (1.2) Z b Gn(x) ηn(dx)= ηn(dx) ηn(dx′)Gn(x′) for some collection ofb likelihoodR functions Gn = gn( ,yn). Replacing these · d1 functions by some ]0, 1]-valued potential function Gn on R , we obtain the conditional distributions of a particle absorption model Xn′ with free evolution transitions Mn and killing rate (1 Gn). More precisely, if T stands for the killing time of the process, we have− that (1.3) η = Law(X′ T >n) and η = Law(X′ T >n). n n| n+1 n+1| These nonabsorption conditional distributions arise in the analysis of con- finement processes,b as well as in computational physics with the numerical solving of Schr¨odinger ground state energies (see, e.g., [11, 13, 28]). Another important class of measures arising in particle physics and stochas- tic optimization problems is the class of Boltzmann distributions, also known CONCENTRATION INEQUALITIES FOR MEAN FIELD PARTICLE MODELS 3 as the Gibbs measure, defined by 1 βnV (x) (1.4) ηn(dx)= e− λ(dx), Zn with some reference probability measure λ, some inverse temperature pa- rameter and some nonnegative potential energy function V on some state space E. The normalizing constant Zn is sometimes called “partition func- tion” or the “free energy.” In sequential Monte Carlo methodology, as well as in operation research literature, these multiplicative formulae are often used to compute rare events probabilities, as well as the cardinality or the volume of some complex state spaces. Further details on these stochastic techniques can be found in the series of articles [5, 29, 30]. To fix the ideas we can consider the uniform measure on some finite set E. Surprisingly, this flow of measures also satisfies the above nonlinear evolution equation, as soon as ηn is an invariant measure of Mn, for each n 1, and the potential functions G are chosen of the following form: G = exp≥ ((β β )V ). The partition n n n+1 − n functions can also be computed in terms of the flow of measures (ηp)0 p<n using the easy-to-check multiplicative formula, ≤ Zn = ηp(dx)Gp(x), 0 p<n ≤Y Z as soon as β0 = 0. In statistical mechanics literature, the above formula is sometimes called the Jarzynski or the Crooks equality [8, 19, 20]. Notice that βn+1V the stochastic models discussed above remain valid if we replace e− by any collection of functions gn+1 s.t. gn+1 = gn Gn = 0 p n Gp, for some × ≤ ≤ potential functions Gn, with n 0. Further details on this model, with sev- eral worked-out applications on≥ concrete hidden MarkovQ chain problems and Bayesian inference can be found in the article [10] dedicated to sequential Monte Carlo technology. All of the models discussed above can be abstracted in a single probabilistic model. The latter is often called a Feynman–Kac model, and it will be presented in some details in Section 2.1. The mean field-type interacting particle system associated with equa- tion (1.1) relies on the fact that the one-step mappings can be rewritten in the following form: (1.5) Φn+1(ηn)= ηnKn+1,ηn for some collection of Markov kernels Kn+1,µ indexed by the time parameter n 0, and the set of measures µ on the space E . We already mention that ≥ n the choice of the Markov transitions Kn,η is not unique. Several examples are presented in Section 2 in the context of Feynman–Kac semigroups or McKean–Vlasov diffusion-type models. To fix the ideas, we can choose el- ementary transitions Kn,η(x,dx′)=Φn(η)(dx′) that do not depend on the state variable x. 4 P. DEL MORAL AND E. RIO In the literature on mean field particle models, the transitions Kn,η are called a choice of McKean transitions. These models provide a natural inter- pretation of the flow of measures ηn as the laws of the time inhomogeneous Markov chain Xn with elementary transitions P (Xn dx Xn 1)= Kn,ηn−1 (Xn 1, dx) with ηn 1 = Law(Xn 1) ∈ | − − − − and starting with some initial random variable with distribution η0 = Law(X0). The Markov chain Xn can be thought of as a perfect sampling algorithm. For a thorough description of these discrete generation and nonlinear McKean- type models, we refer the reader to [9]. In the further development of the article, we always assume that the mappings i N i (xn)1 i N En K N (xn, An+1) n+1,1/N Pj=1 δ j ≤ ≤ ∈ 7→ xn are N -measurable, for any n 0, N 1, and 1 i N, and any mea- En⊗ ≥ ≥ ≤ ≤ surable subset An+1 En+1. In this situation, the mean field particle inter- ⊂ N pretation of this nonlinear measure valued model is an En -valued Markov (N) (N,i) chain ξn = (ξn )1 i N , with elementary transitions defined as ≤ ≤ N (N) (N) (N,i) i P(ξ dx )= K N (ξ , dx ) n+1 ∈ |Fn n+1,ηn n i=1 (1.6) Y N N 1 with ηn := δ (N,j) . N ξn Xj=1 N In the above displayed formula, n stands for the σ-field generated by (N) F 1 N the random sequence (ξp )0 p n, and dx = dx dx stands for an ≤ ≤ 1 ×···×N N infinitesimal neighborhood of a point x = (x ,...,x ) En . The initial sys- (N) ∈ tem ξ0 consists of N independent and identically distributed random vari- ables with common law η0. As usual, to simplify the presentation, when there is no possible confusion we suppress the parameter N, so that we write ξn i (N) (N,i) and ξn instead of ξn and ξn .
Load more

Recommended publications
  • Ruelle Operator for Continuous Potentials and DLR-Gibbs Measures
    Ruelle Operator for Continuous Potentials and DLR-Gibbs Measures Leandro Cioletti Artur O. Lopes Departamento de Matem´atica - UnB Departamento de Matem´atica - UFRGS 70910-900, Bras´ılia, Brazil 91509-900, Porto Alegre, Brazil [email protected] [email protected] Manuel Stadlbauer Departamento de Matem´atica - UFRJ 21941-909, Rio de Janeiro, Brazil [email protected] October 25, 2019 Abstract In this work we study the Ruelle Operator associated to a continuous potential defined on a countable product of a compact metric space. We prove a generaliza- tion of Bowen’s criterion for the uniqueness of the eigenmeasures. One of the main results of the article is to show that a probability is DLR-Gibbs (associated to a continuous translation invariant specification), if and only if, is an eigenprobability for the transpose of the Ruelle operator. Bounded extensions of the Ruelle operator to the Lebesgue space of integrable functions, with respect to the eigenmeasures, are studied and the problem of exis- tence of maximal positive eigenfunctions for them is considered. One of our main results in this direction is the existence of such positive eigenfunctions for Bowen’s potential in the setting of a compact and metric alphabet. We also present a version of Dobrushin’s Theorem in the setting of Thermodynamic Formalism. Keywords: Thermodynamic Formalism, Ruelle operator, continuous potentials, Eigenfunc- tions, Equilibrium states, DLR-Gibbs Measures, uncountable alphabet. arXiv:1608.03881v6 [math.DS] 24 Oct 2019 MSC2010: 37D35, 28Dxx, 37C30. 1 Introduction The classical Ruelle operator needs no introduction and nowadays is a key concept of Thermodynamic Formalism.
    [Show full text]
  • An Infinite Dimensional Central Limit Theorem for Correlated Martingales
    Ann. I. H. Poincaré – PR 40 (2004) 167–196 www.elsevier.com/locate/anihpb An infinite dimensional central limit theorem for correlated martingales Ilie Grigorescu 1 Department of Mathematics, University of Miami, 1365 Memorial Drive, Ungar Building, Room 525, Coral Gables, FL 33146, USA Received 6 March 2003; accepted 27 April 2003 Abstract The paper derives a functional central limit theorem for the empirical distributions of a system of strongly correlated continuous martingales at the level of the full trajectory space. We provide a general class of functionals for which the weak convergence to a centered Gaussian random field takes place. An explicit formula for the covariance is established and a characterization of the limit is given in terms of an inductive system of SPDEs. We also show a density theorem for a Sobolev- type class of functionals on the space of continuous functions. 2003 Elsevier SAS. All rights reserved. Résumé L’article présent dérive d’un théorème limite centrale fonctionnelle au niveau de l’espace de toutes les trajectoires continues pour les distributions empiriques d’un système de martingales fortement corrélées. Nous fournissons une classe générale de fonctions pour lesquelles est établie la convergence faible vers un champ aléatoire gaussien centré. Une formule explicite pour la covariance est determinée et on offre une charactérisation de la limite à l’aide d’un système inductif d’équations aux dérivées partielles stochastiques. On démontre également que l’espace de fonctions pour lesquelles le champ des fluctuations converge est dense dans une classe de fonctionnelles de type Sobolev sur l’espace des trajectoires continues.
    [Show full text]
  • Kalman and Particle Filtering
    Abstract: The Kalman and Particle filters are algorithms that recursively update an estimate of the state and find the innovations driving a stochastic process given a sequence of observations. The Kalman filter accomplishes this goal by linear projections, while the Particle filter does so by a sequential Monte Carlo method. With the state estimates, we can forecast and smooth the stochastic process. With the innovations, we can estimate the parameters of the model. The article discusses how to set a dynamic model in a state-space form, derives the Kalman and Particle filters, and explains how to use them for estimation. Kalman and Particle Filtering The Kalman and Particle filters are algorithms that recursively update an estimate of the state and find the innovations driving a stochastic process given a sequence of observations. The Kalman filter accomplishes this goal by linear projections, while the Particle filter does so by a sequential Monte Carlo method. Since both filters start with a state-space representation of the stochastic processes of interest, section 1 presents the state-space form of a dynamic model. Then, section 2 intro- duces the Kalman filter and section 3 develops the Particle filter. For extended expositions of this material, see Doucet, de Freitas, and Gordon (2001), Durbin and Koopman (2001), and Ljungqvist and Sargent (2004). 1. The state-space representation of a dynamic model A large class of dynamic models can be represented by a state-space form: Xt+1 = ϕ (Xt,Wt+1; γ) (1) Yt = g (Xt,Vt; γ) . (2) This representation handles a stochastic process by finding three objects: a vector that l describes the position of the system (a state, Xt X R ) and two functions, one mapping ∈ ⊂ 1 the state today into the state tomorrow (the transition equation, (1)) and one mapping the state into observables, Yt (the measurement equation, (2)).
    [Show full text]
  • Stochastic Differential Equations with Variational Wishart Diffusions
    Stochastic Differential Equations with Variational Wishart Diffusions Martin Jørgensen 1 Marc Peter Deisenroth 2 Hugh Salimbeni 3 Abstract Why model the process noise? Assume that in the example above, the two states represent meteorological measure- We present a Bayesian non-parametric way of in- ments: rainfall and wind speed. Both are influenced by ferring stochastic differential equations for both confounders, such as atmospheric pressure, which are not regression tasks and continuous-time dynamical measured directly. This effect can in the case of the model modelling. The work has high emphasis on the in (1) only be modelled through the diffusion . Moreover, stochastic part of the differential equation, also wind and rain may not correlate identically for all states of known as the diffusion, and modelling it by means the confounders. of Wishart processes. Further, we present a semi- parametric approach that allows the framework Dynamical modelling with focus in the noise-term is not to scale to high dimensions. This successfully a new area of research. The most prominent one is the lead us onto how to model both latent and auto- Auto-Regressive Conditional Heteroskedasticity (ARCH) regressive temporal systems with conditional het- model (Engle, 1982), which is central to scientific fields eroskedastic noise. We provide experimental ev- like econometrics, climate science and meteorology. The idence that modelling diffusion often improves approach in these models is to estimate large process noise performance and that this randomness in the dif- when the system is exposed to a shock, i.e. an unforeseen ferential equation can be essential to avoid over- significant change in states.
    [Show full text]
  • Applying Particle Filtering in Both Aggregated and Age-Structured Population Compartmental Models of Pre-Vaccination Measles
    bioRxiv preprint doi: https://doi.org/10.1101/340661; this version posted June 6, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY 4.0 International license. Applying particle filtering in both aggregated and age-structured population compartmental models of pre-vaccination measles Xiaoyan Li1*, Alexander Doroshenko2, Nathaniel D. Osgood1 1 Department of Computer Science, University of Saskatchewan, Saskatoon, Saskatchewan, Canada 2 Department of Medicine, Division of Preventive Medicine, University of Alberta, Edmonton, Alberta, Canada * [email protected] Abstract Measles is a highly transmissible disease and is one of the leading causes of death among young children under 5 globally. While the use of ongoing surveillance data and { recently { dynamic models offer insight on measles dynamics, both suffer notable shortcomings when applied to measles outbreak prediction. In this paper, we apply the Sequential Monte Carlo approach of particle filtering, incorporating reported measles incidence for Saskatchewan during the pre-vaccination era, using an adaptation of a previously contributed measles compartmental model. To secure further insight, we also perform particle filtering on an age structured adaptation of the model in which the population is divided into two interacting age groups { children and adults. The results indicate that, when used with a suitable dynamic model, particle filtering can offer high predictive capacity for measles dynamics and outbreak occurrence in a low vaccination context. We have investigated five particle filtering models in this project. Based on the most competitive model as evaluated by predictive accuracy, we have performed prediction and outbreak classification analysis.
    [Show full text]
  • Pdf File of Second Edition, January 2018
    Probability on Graphs Random Processes on Graphs and Lattices Second Edition, 2018 GEOFFREY GRIMMETT Statistical Laboratory University of Cambridge copyright Geoffrey Grimmett Geoffrey Grimmett Statistical Laboratory Centre for Mathematical Sciences University of Cambridge Wilberforce Road Cambridge CB3 0WB United Kingdom 2000 MSC: (Primary) 60K35, 82B20, (Secondary) 05C80, 82B43, 82C22 With 56 Figures copyright Geoffrey Grimmett Contents Preface ix 1 Random Walks on Graphs 1 1.1 Random Walks and Reversible Markov Chains 1 1.2 Electrical Networks 3 1.3 FlowsandEnergy 8 1.4 RecurrenceandResistance 11 1.5 Polya's Theorem 14 1.6 GraphTheory 16 1.7 Exercises 18 2 Uniform Spanning Tree 21 2.1 De®nition 21 2.2 Wilson's Algorithm 23 2.3 Weak Limits on Lattices 28 2.4 Uniform Forest 31 2.5 Schramm±LownerEvolutionsÈ 32 2.6 Exercises 36 3 Percolation and Self-Avoiding Walks 39 3.1 PercolationandPhaseTransition 39 3.2 Self-Avoiding Walks 42 3.3 ConnectiveConstantoftheHexagonalLattice 45 3.4 CoupledPercolation 53 3.5 Oriented Percolation 53 3.6 Exercises 56 4 Association and In¯uence 59 4.1 Holley Inequality 59 4.2 FKG Inequality 62 4.3 BK Inequalitycopyright Geoffrey Grimmett63 vi Contents 4.4 HoeffdingInequality 65 4.5 In¯uenceforProductMeasures 67 4.6 ProofsofIn¯uenceTheorems 72 4.7 Russo'sFormulaandSharpThresholds 80 4.8 Exercises 83 5 Further Percolation 86 5.1 Subcritical Phase 86 5.2 Supercritical Phase 90 5.3 UniquenessoftheIn®niteCluster 96 5.4 Phase Transition 99 5.5 OpenPathsinAnnuli 103 5.6 The Critical Probability in Two Dimensions 107
    [Show full text]
  • The Dobrushin Comparison Theorem Is a Powerful Tool to Bound The
    COMPARISON THEOREMS FOR GIBBS MEASURES∗ BY PATRICK REBESCHINI AND RAMON VAN HANDEL Princeton University The Dobrushin comparison theorem is a powerful tool to bound the dif- ference between the marginals of high-dimensional probability distributions in terms of their local specifications. Originally introduced to prove unique- ness and decay of correlations of Gibbs measures, it has been widely used in statistical mechanics as well as in the analysis of algorithms on random fields and interacting Markov chains. However, the classical comparison theorem requires validity of the Dobrushin uniqueness criterion, essentially restricting its applicability in most models to a small subset of the natural parameter space. In this paper we develop generalized Dobrushin comparison theorems in terms of influences between blocks of sites, in the spirit of Dobrushin- Shlosman and Weitz, that substantially extend the range of applicability of the classical comparison theorem. Our proofs are based on the analysis of an associated family of Markov chains. We develop in detail an application of our main results to the analysis of sequential Monte Carlo algorithms for filtering in high dimension. CONTENTS 1 Introduction . .2 2 Main Results . .4 2.1 Setting and notation . .4 2.2 Main result . .6 2.3 The classical comparison theorem . .8 2.4 Alternative assumptions . .9 2.5 A one-sided comparison theorem . 11 3 Proofs . 12 3.1 General comparison principle . 13 3.2 Gibbs samplers . 14 3.3 Proof of Theorem 2.4................................. 18 3.4 Proof of Corollary 2.8................................. 23 3.5 Proof of Theorem 2.12................................ 25 4 Application: Particle Filters .
    [Show full text]
  • Mean Field Methods for Classification with Gaussian Processes
    Mean field methods for classification with Gaussian processes Manfred Opper Neural Computing Research Group Division of Electronic Engineering and Computer Science Aston University Birmingham B4 7ET, UK. opperm~aston.ac.uk Ole Winther Theoretical Physics II, Lund University, S6lvegatan 14 A S-223 62 Lund, Sweden CONNECT, The Niels Bohr Institute, University of Copenhagen Blegdamsvej 17, 2100 Copenhagen 0, Denmark winther~thep.lu.se Abstract We discuss the application of TAP mean field methods known from the Statistical Mechanics of disordered systems to Bayesian classifi­ cation models with Gaussian processes. In contrast to previous ap­ proaches, no knowledge about the distribution of inputs is needed. Simulation results for the Sonar data set are given. 1 Modeling with Gaussian Processes Bayesian models which are based on Gaussian prior distributions on function spaces are promising non-parametric statistical tools. They have been recently introduced into the Neural Computation community (Neal 1996, Williams & Rasmussen 1996, Mackay 1997). To give their basic definition, we assume that the likelihood of the output or target variable T for a given input s E RN can be written in the form p(Tlh(s)) where h : RN --+ R is a priori assumed to be a Gaussian random field. If we assume fields with zero prior mean, the statistics of h is entirely defined by the second order correlations C(s, S') == E[h(s)h(S')], where E denotes expectations 310 M Opper and 0. Winther with respect to the prior. Interesting examples are C(s, s') (1) C(s, s') (2) The choice (1) can be motivated as a limit of a two-layered neural network with infinitely many hidden units with factorizable input-hidden weight priors (Williams 1997).
    [Show full text]
  • Particle Filter Based Monitoring and Prediction of Spatiotemporal Corrosion Using Successive Measurements of Structural Responses
    sensors Article Particle Filter Based Monitoring and Prediction of Spatiotemporal Corrosion Using Successive Measurements of Structural Responses Sang-ri Yi and Junho Song * Department of Civil and Environmental Engineering, Seoul National University, Seoul 08826, Korea; [email protected] * Correspondence: [email protected]; Tel.: +82-2-880-7369 Received: 11 October 2018; Accepted: 10 November 2018; Published: 13 November 2018 Abstract: Prediction of structural deterioration is a challenging task due to various uncertainties and temporal changes in the environmental conditions, measurement noises as well as errors of mathematical models used for predicting the deterioration progress. Monitoring of deterioration progress is also challenging even with successive measurements, especially when only indirect measurements such as structural responses are available. Recent developments of Bayesian filters and Bayesian inversion methods make it possible to address these challenges through probabilistic assimilation of successive measurement data and deterioration progress models. To this end, this paper proposes a new framework to monitor and predict the spatiotemporal progress of structural deterioration using successive, indirect and noisy measurements. The framework adopts particle filter for the purpose of real-time monitoring and prediction of corrosion states and probabilistic inference of uncertain and/or time-varying parameters in the corrosion progress model. In order to infer deterioration states from sparse indirect inspection data, for example structural responses at sensor locations, a Bayesian inversion method is integrated with the particle filter. The dimension of a continuous domain is reduced by the use of basis functions of truncated Karhunen-Loève expansion. The proposed framework is demonstrated and successfully tested by numerical experiments of reinforcement bar and steel plates subject to corrosion.
    [Show full text]
  • Dynamic Detection of Change Points in Long Time Series
    AISM (2007) 59: 349–366 DOI 10.1007/s10463-006-0053-9 Nicolas Chopin Dynamic detection of change points in long time series Received: 23 March 2005 / Revised: 8 September 2005 / Published online: 17 June 2006 © The Institute of Statistical Mathematics, Tokyo 2006 Abstract We consider the problem of detecting change points (structural changes) in long sequences of data, whether in a sequential fashion or not, and without assuming prior knowledge of the number of these change points. We reformulate this problem as the Bayesian filtering and smoothing of a non standard state space model. Towards this goal, we build a hybrid algorithm that relies on particle filter- ing and Markov chain Monte Carlo ideas. The approach is illustrated by a GARCH change point model. Keywords Change point models · GARCH models · Markov chain Monte Carlo · Particle filter · Sequential Monte Carlo · State state models 1 Introduction The assumption that an observed time series follows the same fixed stationary model over a very long period is rarely realistic. In economic applications for instance, common sense suggests that the behaviour of economic agents may change abruptly under the effect of economic policy, political events, etc. For example, Mikosch and St˘aric˘a (2003, 2004) point out that GARCH models fit very poorly too long sequences of financial data, say 20 years of daily log-returns of some speculative asset. Despite this, these models remain highly popular, thanks to their forecast ability (at least on short to medium-sized time series) and their elegant simplicity (which facilitates economic interpretation). Against the common trend of build- ing more and more sophisticated stationary models that may spuriously provide a better fit for such long sequences, the aforementioned authors argue that GARCH models remain a good ‘local’ approximation of the behaviour of financial data, N.
    [Show full text]
  • Probability on Graphs Random Processes on Graphs and Lattices
    Probability on Graphs Random Processes on Graphs and Lattices GEOFFREY GRIMMETT Statistical Laboratory University of Cambridge c G. R. Grimmett 1/4/10, 17/11/10, 5/7/12 Geoffrey Grimmett Statistical Laboratory Centre for Mathematical Sciences University of Cambridge Wilberforce Road Cambridge CB3 0WB United Kingdom 2000 MSC: (Primary) 60K35, 82B20, (Secondary) 05C80, 82B43, 82C22 With 44 Figures c G. R. Grimmett 1/4/10, 17/11/10, 5/7/12 Contents Preface ix 1 Random walks on graphs 1 1.1 RandomwalksandreversibleMarkovchains 1 1.2 Electrical networks 3 1.3 Flowsandenergy 8 1.4 Recurrenceandresistance 11 1.5 Polya's theorem 14 1.6 Graphtheory 16 1.7 Exercises 18 2 Uniform spanning tree 21 2.1 De®nition 21 2.2 Wilson's algorithm 23 2.3 Weak limits on lattices 28 2.4 Uniform forest 31 2.5 Schramm±LownerevolutionsÈ 32 2.6 Exercises 37 3 Percolation and self-avoiding walk 39 3.1 Percolationandphasetransition 39 3.2 Self-avoiding walks 42 3.3 Coupledpercolation 45 3.4 Orientedpercolation 45 3.5 Exercises 48 4 Association and in¯uence 50 4.1 Holley inequality 50 4.2 FKGinequality 53 4.3 BK inequality 54 4.4 Hoeffdinginequality 56 c G. R. Grimmett 1/4/10, 17/11/10, 5/7/12 vi Contents 4.5 In¯uenceforproductmeasures 58 4.6 Proofsofin¯uencetheorems 63 4.7 Russo'sformulaandsharpthresholds 75 4.8 Exercises 78 5 Further percolation 81 5.1 Subcritical phase 81 5.2 Supercritical phase 86 5.3 Uniquenessofthein®nitecluster 92 5.4 Phase transition 95 5.5 Openpathsinannuli 99 5.6 The critical probability in two dimensions 103 5.7 Cardy's formula 110 5.8 The
    [Show full text]
  • Stochastic Differential Equations with Variational Wishart Diffusions
    Stochastic Differential Equations with Variational Wishart Diffusions Martin Jørgensen 1 Marc Peter Deisenroth 2 Hugh Salimbeni 3 Abstract Why model the process noise? Assume that in the example above, the two states represent meteorological measure- We present a Bayesian non-parametric way of in- ments: rainfall and wind speed. Both are influenced by ferring stochastic differential equations for both confounders, such as atmospheric pressure, which are not regression tasks and continuous-time dynamical measured directly. This effect can in the case of the model modelling. The work has high emphasis on the in (1) only be modelled through the diffusion . Moreover, stochastic part of the differential equation, also wind and rain may not correlate identically for all states of known as the diffusion, and modelling it by means the confounders. of Wishart processes. Further, we present a semi- parametric approach that allows the framework Dynamical modelling with focus in the noise-term is not to scale to high dimensions. This successfully a new area of research. The most prominent one is the lead us onto how to model both latent and auto- Auto-Regressive Conditional Heteroskedasticity (ARCH) regressive temporal systems with conditional het- model (Engle, 1982), which is central to scientific fields eroskedastic noise. We provide experimental ev- like econometrics, climate science and meteorology. The idence that modelling diffusion often improves approach in these models is to estimate large process noise performance and that this randomness in the dif- when the system is exposed to a shock, i.e. an unforeseen ferential equation can be essential to avoid over- significant change in states.
    [Show full text]